By Priyanjana Pramanik, MSc. Despite near-perfect exam scores, large language models falter when real people rely on them for ...